For neural networks, function determines form

Neural Networks

Volumn 6, Issue 7, 1993, Pages 975-990

  Albertini, Francesca ^a,b   Sontag, Eduardo D ^a  

^a RUTGERS UNIVERSITY   (United States)

^b UNIVERSITY OF PADOVA   (Italy)

Author keywords

Control systems; Identification from input output data; Neural networks

Indexed keywords

COMPUTATIONAL METHODS; CONTROL SYSTEMS; DIFFERENTIAL EQUATIONS; IDENTIFICATION (CONTROL SYSTEMS); MATHEMATICAL MODELS; NONLINEAR EQUATIONS;

BLACK BOXES; CONTINUOUS TIME FEEDBACK NEURAL NETWORKS; NONLINEAR ACTIVATION FUNCTION; WEAK GENERICITY ASSUMPTIONS;

NEURAL NETWORKS;

EID: 0027795346     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0893-6080(09)80007-5     Document Type: Article

References (21)

1. Identiflability of discrete-time neural networks
2. Uniqueness of weights for neural networks
3. Finite state automata and simple recurrent networks
4. Absolute stability of global pattern formation and parallel memory storage by competitive neural networks
5. Statistical mechanics of neural nets
6. Theory of the backpropagation neural network
7. Convergent activation dynamics in continuous-time networks
8. Neurons with graded responses have collective computational properties like those of two-state neurons
9. Qualitative analysis of neural networks
10. Static and dynamic error propagation networks with application to speech coding
11. Algebraic theory of sign-linear systems
12. Turning computability with neural nets
13. On the computational power of neural nets
14. Some results on computing with ‘neural nets’
15. On the observability of polynomial systems
16. Single-input observability of continuous-time systems
17. Existence and uniqueness of minimal realizations of nonlinear systems
18. Uniqueness of the weights for minimal feedforward nets with a given input-output map